A Stochastic Approach to the Measurement and Analysis of Leadscrew Drive Kinematic Errors

Abstract

A practical way to achieve machine tool drive system accuracy is through compensation in the position control loop. Various strategies have been developed in which the compensation data are stored either in a computer or by alternative techniques. The approach described in this paper demonstrates how the dominant causes of the inaccuracies in machine tool drive systems can be determined in a simple and straightforward manner and how the compensation data can be generated using a microprocessor-based calibration technique. The proposed method was demonstrated on a Giddings & Lewis Model 10V Numericenter and consists of three major steps: 1) accuracy measurement, 2) error-map representation, and 3) error-map analysis and error cause diagnosis. Accuracy measurement was accomplished by reading both the resolver counts and the output of a laser interferometer which served as a reference (master) scale. The procedure is based on a strategy in which small 0.5 mm (0.020 in.), highly accurate ±2.5 μm ( ± 0.0001 in.) linear displacements of the machine tool table motion are taken as the measurement reference. An ARMA (Autoregressive Moving Average) model of order (16, 15) is the statistically adequate model that fit the data of the angular displacement of the resolver. The Dynamic Data System technique was used for this purpose and it can also be used to represent the error-map function. A spectral analysis of the ARMA model revealed the sources of the kinematic errors.